Methods, control methods, chips, equipment and media for testing automatic control systems

Abstract

This disclosure provides a method, control method, chip, device, and medium for detecting an automatic control system. The method includes: acquiring a perception map of a target scene generated by the automatic control system based on target scene data; wherein the target scene data is collected by a mobile device in the target scene; acquiring a truth map of the target scene based on truth value annotation of the target scene data; and comparing the perception map and the truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene; wherein the indicator set includes indicators associated with the position of the mobile device in the perception map.

Description

Technical Field

[0001] This disclosure relates to the field of automatic control technology, and more specifically, to a method for detecting an automatic control system, a control method for a mobile device, a chip, an electronic device, a mobile device, and a computer-readable storage medium. Background Technology

[0002] For mobile devices such as vehicles, unmanned automated movement can be achieved through a matching automatic control system. Specifically, the automatic control system determines scene information by collecting scene data from the mobile device, thereby enabling movement control. Therefore, the scene perception capability of the automatic control system is crucial for achieving movement control of the mobile device. It is necessary to test the scene perception capability of the automatic control system to improve its perception accuracy and stability based on the test results. Existing detection methods are based on the objects within the perception range and cannot determine the perception capability of the automatic control system for the objects of interest required to complete the movement control of the equipment, thus affecting the effectiveness of the detection. Summary of the Invention

[0003] In view of this, the present disclosure proposes a new technical solution for detecting the scene perception capability of an automatic control system.

[0004] According to a first aspect of the present disclosure, a method for detecting an automatic control system is provided, the method comprising:

[0005] A method is provided for acquiring a perception map of a target scene generated by an automatic control system based on target scene data; wherein the target scene data is collected by a mobile device in the target scene;

[0006] Obtain a truth map of the target scene based on truth labeling of the target scene data;

[0007] By comparing the perception map and the ground truth map based on a set of defined indicators, a detection result reflecting the scene perception capability of the automatic control system for the target scene is obtained; wherein, the set of defined indicators includes indicators associated with the position of the mobile device in the perception map.

[0008] Optionally, the set of indicators includes at least one of local indicators associated with the location and global indicators associated with the location; wherein,

[0009] Local metrics associated with the location include: metrics for detecting the scene perception capability based on a local region of the perception map, wherein the local region is a region determined based on the location of the mobile device in the perception map;

[0010] The global metrics associated with the location include: metrics for the scene perception capability based on the entire area of ​​the perception map and objects associated with the location.

[0011] Optionally, the perception map indicates the passable area of ​​the mobile device in the target scene as perceived by the automatic control system; the truth map indicates the passable area of ​​the mobile device in the target scene determined based on truth labels; and the set of indicators includes indicators associated with the passable area.

[0012] Optionally, the set of indicators is a first set of indicators that matches the first mobile task corresponding to the target scenario;

[0013] The first mobile task is a mobile task in a set of defined tasks. The set of defined tasks also includes a second mobile task. The second set of indicators that matches the second mobile task is different from the first set of indicators that matches the first mobile task.

[0014] Optionally, the target scene data includes multiple frames of data, with the same frame corresponding to the same timestamp and different frames corresponding to different timestamps; the perception map includes multiple perception map segments that correspond one-to-one with the multiple frames of data, and the truth map includes multiple truth map segments that correspond one-to-one with the multiple frames of data.

[0015] The step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result representing the scene perception capability of the automatic control system for the target scene includes:

[0016] Based on a set of indicators, the perceptual map fragments and ground truth map fragments of corresponding data in the same frame are compared to obtain the detection results.

[0017] Optionally, when the first mobile task is a first type of task, the set of indicators includes local indicators of the scene perception capability based on the local area detection of the perception map.

[0018] When the first mobile task is a second type of task, the indicators in the indicator set are global indicators of the scene perception capability based on the entire area of ​​the perception map.

[0019] Optionally, the global indicator is an indicator shared by different mobile tasks, and the step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes:

[0020] Based on the global index, the perception map and the truth map are compared to obtain the first detection result for the automatic control system.

[0021] If the first mobile task is the second type of task, the first detection result is output as the detection result.

[0022] If the first mobile task is a first type of task, determine the local metrics that match the first mobile task;

[0023] Based on the determined local indicators, the perception map and the truth map are compared to obtain a second detection result for the automatic control system.

[0024] The first detection result and the second detection result are output as the detection result.

[0025] Optionally, the global indicators associated with the location include a first global indicator for detecting contour differences in traversable areas, the detection result including the indicator value of the automatic control system for the first global indicator; the step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes:

[0026] A set of sampling rays is emitted from the position of the mobile device in the sensing map to the periphery, and the contour deviation value of the first contour of the passable area indicated by the sensing map is collected through the sampling rays relative to the second contour of the passable area indicated by the ground truth map; wherein the contour deviation value is determined based on the first intersection point of the sampling ray with the first contour and the second intersection point of the sampling ray with the second contour.

[0027] Based on the contour deviation values ​​collected by the set of sampling rays, the index value of the automatic control system for the first global index is determined.

[0028] Optionally, the local indicators associated with the location include a first local indicator for detecting obstacle boundary errors, and the detection result includes the indicator value of the automatic control system for the first local indicator; the step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes:

[0029] Based on the perception map, target obstacles within a first predetermined range of the mobile device are identified;

[0030] Based on each target obstacle, using the true boundary of the target obstacle in the truth map as a reference, the set offset of the perceived boundary of the target obstacle in the perception map relative to the true boundary is determined, and used as the index value of the automatic control system for the first local index.

[0031] Wherein, the set offset is the offset of the visible portion of the perceived boundary within the visible range of the mobile device relative to the real boundary.

[0032] Optionally, the local indicators associated with the location include a second local indicator for detecting the accuracy of the boundary of the target entering the area within the passable area, and the detection result includes the indicator value of the automatic control system for the second local indicator; the step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes:

[0033] Based on the perception map, the target entry area located within a second predetermined range of the mobile device is determined;

[0034] For each target entry area, the passable portion of the ground truth map and the perception map in the extended region of the target entry area is compared to obtain the index value of the automatic control system for the second local index.

[0035] According to a second aspect of this disclosure, a control method for a mobile device is also provided, the control method being implemented by a first automatic control system, the method comprising:

[0036] Acquire the first scene data collected by the first mobile device;

[0037] Based on the first scene data, control the movement of the first mobile device;

[0038] The first automatic control system obtains the values ​​of the parameters of the second automatic control system by adjusting the parameters of the second automatic control system. The adjustment is based on the detection results of the scene perception capability of the second automatic control system for at least one target scene. The detection results are obtained by comparing the perception map and the truth map based on the index set. The index set includes indicators associated with the position of the mobile device in the perception map. The perception map is generated by the second automatic control system based on the second scene data collected by the second mobile device in the target scene. The truth map is obtained by labeling the second scene data with truth values.

[0039] According to a third aspect of this disclosure, a chip is provided according to some embodiments, the chip including:

[0040] Storage unit for storing computer programs; and,

[0041] A processing unit configured to implement the method according to the first and / or second aspects of this disclosure when executing a computer program stored in the storage unit.

[0042] According to a fourth aspect of the present disclosure, an electronic device is provided, the electronic device including a memory and a processor, the memory being configured to store computer instructions, and the processor being configured to invoke the computer instructions from the memory to perform the methods described in the first aspect and / or the second aspect.

[0043] According to a fifth aspect of this disclosure, a mobile device is provided according to some embodiments, which may include a chip according to a third aspect of this disclosure; or, include an electronic device according to a fourth aspect of this disclosure; or, may include:

[0044] processor;

[0045] Memory used to store processor-executable instructions;

[0046] The processor is configured to implement the method according to the first and / or second aspects of this disclosure when executing instructions stored in the memory.

[0047] According to a sixth aspect of this disclosure, a non-volatile computer-readable storage medium is also provided, having stored thereon computer program instructions that, when executed by a processor, implement the above-described method.

[0048] According to the method provided in this disclosure, based on indicators associated with the position of the mobile device in the perception map, the perception map generated by the automatic control system based on target scene data is compared with a truth map obtained by ground truth annotation of the target scene data, thereby obtaining a detection result reflecting the scene perception capability of the automatic control system for the target scene. Here, since the position-associated indicators can obtain a deviation from position-based perception, and the position of the mobile device in the scene is highly correlated with the automatic control system's ability to complete the movement task in the target scene, detecting the automatic control system based on such indicators can focus on the objects of interest required by the automatic control system to complete the device movement control, thereby improving the effectiveness of detection and system optimization based on the detection results.

[0049] Other features and advantages of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0050] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments of the present disclosure and, together with their description, serve to explain the principles of the present disclosure.

[0051] Figure 1 This is a schematic diagram of an intelligent connected system to which the methods provided in the embodiments of this disclosure can be applied;

[0052] Figure 2 It is based on Figure 1 The illustrated embodiment provides a schematic diagram of a mobile device;

[0053] Figure 3 This is a flowchart illustrating a method for detecting an automatic control system according to some embodiments;

[0054] Figure 4 It is a schematic diagram comparing the perception map and the truth map for the first local index based on some examples;

[0055] Figure 5 This is a schematic diagram comparing the perception map and the truth map for the second local indicator, based on some examples.

[0056] Figure 6 This is a schematic diagram comparing a perception map and a ground truth map for a first global metric, based on some examples.

[0057] Figure 7 This is a schematic diagram comparing perception maps and truth maps for the second type of indicators, based on some examples;

[0058] Figure 8 This is a flowchart illustrating a method for detecting an automatic control system according to other embodiments;

[0059] Figure 9 This is a flowchart illustrating a control method for a mobile device according to some embodiments;

[0060] Figure 10 This is a schematic diagram of the chip's structural composition according to some embodiments;

[0061] Figure 11 This is a schematic diagram of the composition structure of an electronic device according to some embodiments;

[0062] Figure 12 This is a schematic diagram of the composition structure of a movable device according to some embodiments. Detailed Implementation

[0063] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0064] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0065] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0066] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0067] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0068] It should be noted that all data acquisition actions in this disclosure were carried out in compliance with the relevant data protection laws and policies of the country where the data is located, and with the authorization of the relevant equipment owner.

[0069] This disclosure relates to a scheme for detecting the scene perception capability of an automatic control system for a mobile device. In embodiments of this disclosure, the automatic control system is a system that perceives the scene in which the mobile device is located based on scene data collected by sensors mounted on the mobile device, determines the passable area of ​​the mobile device within the scene, and then controls the movement of the mobile device within the scene. Taking a vehicle as an example, the automatic control system for a vehicle can be called an autonomous driving system. The automatic control system can be deployed on the mobile device, on a server communicating with the mobile device, or on both the mobile device and the server; no limitation is made herein.

[0070] In technologies related to detecting the scene perception capability of automatic control systems, the main focus is on determining whether the scene perception capability of the automatic control system meets application requirements based on the overall precision, recall, and accuracy of the automatic control system's perception results of scene data. Among these, precision is the proportion of samples that are actually positive out of those that the system identifies as positive; recall is the proportion of all samples that are actually positive out of those that the system correctly identifies as positive; and accuracy is the proportion of samples that the system correctly identifies out of the total number of samples.

[0071] The above-mentioned indicators focus on reflecting the overall perception capability of the automatic control system for the scene, lacking scene-specificity. They cannot reflect the automatic control system's perception capability for the objects of interest required to complete the movement control of the equipment. In other words, they cannot accurately detect the automatic control system, which limits the performance optimization of the automatic control system, affects the perception effect of the automatic control system on the objects of interest, and thus affects the safety of movement control. To address this, embodiments of this disclosure propose a method for detecting automatic control systems to improve the accuracy and effectiveness of detection. Accordingly, embodiments of this disclosure also propose an optimization method for automatic control systems based on the detection results, and a control method for controlling the movement of mobile devices based on the optimized automatic control system.

[0072] Figure 1 An intelligent connected system 100 that can apply the methods provided in the embodiments of this disclosure is illustrated. Figure 1 As shown, the intelligent connected system 100 may include: a mobile device 101, a server 102, and a user terminal 103.

[0073] In some examples, the mobile device 101 can be a vehicle with autonomous driving capabilities, a robot capable of autonomous movement, etc. Autonomous driving, also known as driverless or intelligent driving, refers to a vehicle with autonomous driving capabilities that can perform driving tasks such as environmental perception, decision-making, planning, and control execution. The levels of autonomous driving can refer to the vehicle intelligence classification standards established by the Society of Automotive Engineers (SAE), for example, L0 is manual driving, L1 is driver assistance, L2 is partial autonomous driving, L3 is conditional autonomous driving, L4 is highly automated driving, and L5 is fully automated driving. The above classification of autonomous driving levels is merely an example, and this disclosure does not limit the classification standards and levels of autonomous driving.

[0074] In some examples, server 102 can be a single server or a distributed server cluster consisting of multiple servers, and its deployment can include local servers and / or cloud servers. Server 102 can communicate with mobile device 101 and / or user terminal 103 via a communication network, providing various services to mobile device 101 and / or user terminal 103. For example, the server can receive sensing data sent by mobile device 101, and provide services such as high-precision maps, data analysis, and decision planning to mobile device 101. Alternatively, the server can receive query commands or control commands sent by user terminal 102, providing corresponding services to the user.

[0075] In some examples, the user terminal 103 can be any form of electronic device that provides services to the user, such as a personal computer, laptop, smart tablet, smartphone, smart wearable device, etc. The user can interact with the mobile device or server through the human-computer interaction terminal configured on the mobile device 101, or through the user terminal 103. For example, the user terminal can query the status and / or parameters of the mobile device, or control the mobile device to perform set tasks and / or modify configuration parameters, etc. The user terminal 103 runs an application based on the intelligent network system to achieve interaction with the mobile device or server. This application can be a local application, a web application, or a mini-program, etc., without limitation.

[0076] In some examples, the aforementioned application running on user terminal 103 can provide authentication or authorization services to users. Users who are successfully authenticated and granted the corresponding permissions can query and / or control the mobile device within the scope of the granted permissions.

[0077] The mobile device 101, server 102, and user terminal 103 can communicate via a communication link provided by communication network 104. This communication network 104 can include one or more networks of any type, such as the Internet, Local Area Network (LAN), Wide Area Network (WAN), Virtual Private Network (VPN), Public Switched Telephone Network (PSTN), satellite communication network, Wi-Fi, 2G, 3G, 4G, 5G, 6G, NB-IoT, eMTC, infrared, Bluetooth, NFC, or a combination of these networks. The communication networks between mobile device 101 and server 102, between user terminal 103 and server 102, and between user terminal 103 and mobile device 101 can be the same or different.

[0078] It should be noted that, Figure 1 The structure of the intelligent connected system 100 shown is merely illustrative. The intelligent connected system in this embodiment is not limited to the above structure and may include more or fewer devices as needed, and the devices may be combined or separated. For example, the intelligent connected system may not include... Figure 1 The user terminal in the middle.

[0079] The method provided in this disclosure can be implemented by the mobile device 101, the server 102, or both. Furthermore, those skilled in the art should understand that the detection automatic control system method and optimization method provided in this disclosure can also be implemented by other devices with data processing capabilities, independent of the intelligent network system; simply importing the scene data collected by the mobile device into that device is sufficient, and no limitation is made herein.

[0080] Figure 2 It is based on Figure 1 The illustrated embodiment provides a schematic diagram of a mobile device 101. As shown... Figure 2 As shown, the mobile device 101 may include a sensing component 1011, a computing platform 1012, an execution component 1013, etc. The sensing component 1011, the computing platform 1012, and the execution component 1013 may be connected via a bus or other means.

[0081] In some examples, the sensing component 1011 can be used to collect information about the mobile device itself or its external environment. The sensing component 1011 may include a visual sensing unit and a motion sensing unit. The visual sensing unit may include one or more cameras. The motion sensing unit may include a wheel speedometer and / or an inertial measurement unit (IMU). In other examples, the sensing component 1011 may also include radar, a positioning and navigation unit, etc., without limitation. The radar may include at least one of lidar, millimeter-wave radar, ultrasonic radar, or other radar types. The wheel speedometer can be of any type, such as a magnetoelectric wheel speedometer, an photoelectric wheel speedometer, a mechanical wheel speedometer, a Hall effect wheel speedometer, or a visual wheel speedometer. The positioning and navigation unit may include at least one of a GPS system, a BeiDou system, or other global positioning systems.

[0082] In some examples, the computing platform 1012 may include a computing-capable device for processing the sensing information collected by the sensing component 1011 to obtain control information, and sending corresponding control commands to the execution component 1013 to cause the execution component 1013 to perform corresponding actions, thereby realizing the control of the mobile device 101. For example, the computing platform 1012 can perform Simultaneous Localization and Mapping (SLAM), path planning, and behavior decision-making on the mobile device, thereby realizing autonomous control of the mobile device. The computing platform 1012 may include at least one processor and at least one memory, and each processor may execute instructions stored in the memory individually or jointly to implement the methods provided in the embodiments of this disclosure. The processor in the embodiments of this disclosure may include at least one of a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a Field Programmable Gate Array (FPGA), a System on Chip (SOC), an Application Specific Integrated Circuit (ASIC), a Microcontroller Unit (MCU), or other processors. Memory can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. In addition to storing instructions, memory can also store data, such as high-definition maps, path information, and data on the location, direction, and speed of mobile devices. The data stored in memory can be accessed and used by the processor.

[0083] In some examples, the computing platform of a mobile device can perform computing tasks independently or communicate with a server to complete computing tasks. For example, the computing platform of a mobile device can cooperate with a server to complete the corresponding computing tasks.

[0084] The computing platform 1012 can be located in the mobile device 101. Some or all of the computing platform 1012 can also be located in the server corresponding to the mobile device. For example, some functions of the computing platform 1012 with high real-time requirements can be located in the mobile device, while other functions with low real-time requirements can be located in the server corresponding to the mobile device.

[0085] In some examples, the execution component 1013 is used to perform corresponding actions based on the control of the computing platform 1012, enabling the mobile device 101 to complete the movement task. The execution component 1013 may include, for example, a power component, a braking component, a transmission component, a steering component, etc.

[0086] It should be noted that, Figure 2 The structure of the mobile device 101 shown is merely illustrative. The mobile device in this embodiment is not limited to the above structure and may include more or fewer components as needed. The device may also be combined or disassembled. For example, the mobile device may not include the aforementioned computing platform. Furthermore, the mobile device may also include communication components, interface components, multimedia components, input components, output components, etc.

[0087] The following is combined with Figure 1 The systems shown illustrate various embodiments of this disclosure.

[0088] <First Embodiment>

[0089] Figure 3 A flowchart illustrating a method for an automatic control system for detection according to some embodiments is shown.

[0090] like Figure 3 As shown, the method of this disclosure embodiment may include the following steps S310 to S330:

[0091] Step S310: Obtain the perception map of the target scene generated by the automatic control system based on the target scene data.

[0092] In this embodiment, the target scene data is collected by the mobile device in the target scene where the mobile task is performed. This mobile task can be any task that the automatic control system can perform. The mobile task is related to the target scene; for example, a vehicle may perform a parking task in a parking lot scene, or a driving task or lane-changing task in a road scene.

[0093] Mobile devices can acquire target scene data through their onboard sensing components. These sensing components include at least one of a camera and radar. Accordingly, the target scene data can include at least one of image data acquired by a camera and point cloud data acquired by radar.

[0094] Target scene data refers to data that provides feedback on the content of the target scene. It can be used to determine the scene element information of the target scene, which includes static element information and dynamic element information.

[0095] Target scene data can include multiple frames of data, in order to Figure 2Taking the mobile device 101 as an example, the sensing component 1011 of the mobile device 101 can collect data at a set sampling frequency. Data in the same frame corresponds to the same timestamp, and data in different frames corresponds to different timestamps. The timestamp is used to mark the sampling time of the data. It can be understood that data in the same frame corresponding to the same timestamp does not mean that the sampling time of these data is strictly aligned. In practical applications, a time window for alignment can be set. That is to say, data collected within a time window can be considered to correspond to the same timestamp.

[0096] In some examples, since the movement of the mobile device is limited, the scene data collected by the sensing component in two sampling sessions will not affect the positioning accuracy. The movement includes both distance and angle. For example, in a parking lot scene, the scene elements are mainly parking spaces and pillars. If the vehicle movement is limited, the scene elements within the radiation range of the sensing component 1011 will remain largely unchanged. Therefore, keyframe data can be selected based on the movement of the mobile device 101 in the scene for processing according to the method of this embodiment, without having to process every frame of data collected, thus reducing the amount of data processing. For example, keyframe data can be selected from the data collected by the mobile device based on the setting that the movement of the mobile device between two adjacent keyframes should be greater than or equal to a set threshold.

[0097] A perceptual map can include multiple perceptual map segments, which are generated based on single-frame data, and different perceptual map segments are generated based on different single-frame data. A perceptual map can also be a stitched map formed by stitching together multiple perceptual map segments.

[0098] In this embodiment, the perceived map can be a bird's-eye view (BEV) map, also known as a bird's-eye view. Depending on the detection requirements, this map can be a two-dimensional view or a three-dimensional view with height information. The bird's-eye view can present scene information in a simple way, facilitating comparison between the perceived map and the ground truth map.

[0099] An automatic control system can be equipped with at least one model. Data processing, scene element detection, and passable area (freespace) detection are performed through at least one model to generate a perception map, which will not be elaborated here.

[0100] Step S320: Obtain the ground truth map of the target scene based on ground truth annotation of the scene data.

[0101] In this embodiment, the truth value of the target scene data can be generated based on truth value production methods known to those skilled in the art. That is, the scene elements represented by the target scene data are labeled with truth values ​​to obtain a truth map. Truth value labeling can be fully automated based on the model, or it can be done manually with labeling or calibration, and there is no limitation here.

[0102] Truth maps and perception maps share the same structural form for ease of comparison. For example, when a perception map comprises multiple perception map segments, a truth map also comprises multiple truth map segments. Furthermore, when a perception map is a stitched map, a truth map is also a stitched map. And again, when a perception map is a bird's-eye view, a truth map is also a bird's-eye view, and so on.

[0103] Step S330: Based on the set index set, compare the perception map and the truth map to obtain the detection results reflecting the scene perception capability of the automatic control system for the target scene.

[0104] In this embodiment, the indicator set includes at least one indicator. At least a portion of the indicators in the indicator set are indicators associated with the location of the mobile device in the perception map. In this disclosure, indicators associated with the location of the mobile device in the perception map are referred to as first-type indicators, while indicators unrelated to the location are referred to as second-type indicators. In some examples, in addition to first-type indicators, the indicator set may also include second-type indicators, which is not limited here.

[0105] In this embodiment, the location association can be reflected in the fact that the map area of ​​interest of the indicator is related to the location of the mobile device in the perceived map.

[0106] In this embodiment, the correlation with location can also be seen in that the determination of the indicator value is related to the location of the mobile device in the perception map. That is, the indicator value is determined based on objects related to the location of the mobile device in the perception map.

[0107] Since the location of the mobile device in the target scene is highly correlated with the automatic control system's completion of the movement task, the first type of indicator can also be an indicator associated with the movement task corresponding to the target scene.

[0108] Besides categorizing indicators into Category I and Category II based on location correlation, indicators can also be classified into global and local indicators based on their coverage area. Local indicators measure the scene perception capability of an automatic control system for detecting objects within a local area of ​​a perception map. These indicators provide refined detection of the perception capability of the automatic control system in the local area it needs to focus on during motion tasks. Global indicators, on the other hand, measure the scene perception capability of an automatic control system for detecting objects across the entire area of ​​the perception map. These indicators measure the overall accuracy of the automatic control system's perception of the target environment, supporting optimization of the automatic control system from the perspective of overall perception capability.

[0109] The first category of indicators in the indicator set can be global indicators, local indicators, or both. The second category of indicators in the indicator set can also include at least one of the local and global indicators.

[0110] In some examples, the first type of indicators in the set of indicators includes local indicators. That is, the set of indicators includes location-associated local indicators, the local area of ​​concern of which is the area within a set range of the mobile device. For the safety of the automatic control system in completing its movement task in the target scenario, the perception capability of the automatic control system in the vicinity of the mobile device is crucial. Therefore, by using location-associated local indicators, the perception capability of the automatic control system in the vicinity of the device's movement control can be determined, avoiding the equalization of detection results by other areas with low correlation. This enables refined detection of the automatic control system, which is beneficial to improving the effectiveness of the detection results in optimizing the performance of the automatic control system, thereby improving the safety of the automatic control system in device movement control.

[0111] Furthermore, since determining the traversable areas of mobile devices within a scene by perceiving the scene is a crucial part of the automatic control system's ability to safely control the movement of mobile devices within that scene, in some examples, the scene perception capability of the automatic control system can be tested based on its detection of traversable areas. In other words, the indicator set can include indicators associated with traversable areas to focus on the effectiveness of traversable area detection. Specifically, the perception map generated by the automatic control system based on target scene data can indicate the traversable areas of the mobile devices perceived by the automatic control system in the target scene, while the truth map can indicate the traversable areas of the mobile devices in the target scene determined based on truth map annotations. Thus, by comparing the perception map and the truth map, the indicator values ​​associated with traversable areas can be determined, thereby enabling the testing of the automatic control system's scene perception capability from the perspective of determining traversable areas.

[0112] In this example, the indicators in the indicator set that are associated with the passable area can have overlap with the first type of indicators. That is, an indicator in the indicator set can be either a first type of indicator or an indicator associated with the passable area. For example, all indicators in the indicator set can be indicators associated with the passable area, without any limitation.

[0113] For the detection of passable areas, since the accuracy of passable area detection depends on the accuracy of obstacle boundary detection, in some examples, the first type of metric may include a first local metric for detecting obstacle boundary errors. This first local metric is also an metric associated with the passable area, so as to determine the automatic control system's ability to detect both obstacle boundaries and passable areas using the first type of metric. In this example, the detection result includes the index value of the automatic control system for this first local metric.

[0114] In this example, step S330, which compares the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene, may include the following steps: based on the perception map, determining the target obstacles within a first set range of the mobile device; and, based on each target obstacle, using the real boundary of the target obstacle in the ground truth map as a reference, determining the set offset of the perception boundary of the target obstacle in the perception map relative to the real boundary, as the indicator value of the automatic control system for the first local indicator.

[0115] In this example, the first set range can be represented by a distance relative to the mobile device, or by a square area of ​​a certain size with the mobile device as the geometric center. The distance or the size of the square area can be set according to the target scene or the mobile task, and is not limited here.

[0116] In this example, the target obstacle can be any obstacle within a first set range, or it can be an obstacle of a set type based on the target scene or movement task. That is, for different target scenes or movement tasks, the obstacle type detected as the first local indicator can be the same or different. For example, if the target scene is a parking lot and the movement task is a parking task, the target obstacles could be other vehicles, pillars, etc., within the first set range. Another example is if the target scene is a road scene and the movement task is a lane-changing task, the target obstacles could be other vehicles, pedestrians, etc., within the first set range.

[0117] In this example, the set offset may include at least one of the following: average offset, maximum offset, minimum offset, offset along a set direction, etc.

[0118] Furthermore, the setting direction can be determined based on the visible range of the mobile device in the sensing map. In other words, the setting offset can be the offset of the visible portion of the sensing boundary within the mobile device's visible range relative to the actual boundary. Here, since the visible portion of the sensing boundary is the part that affects the movement of the mobile device, the setting offset can only be the offset of the visible portion. For example, the setting offset can be the maximum offset, average offset, or offset in the setting direction within the visible portion. This not only improves the effectiveness of the indicator in detecting the automatic control system's ability to perform movement tasks but also reduces data processing volume and improves detection efficiency.

[0119] Figure 4 A portion of a perception map for a parking task is schematically illustrated. This perception map is generated by the automatic control system based on scene data collected by vehicle X1. Other vehicles X2 in the perception map are target obstacles located within a first predetermined range of vehicle X1. When determining the predetermined offset of the perception boundary W2 of vehicle X2 in the perception map relative to the real boundary W2′ of vehicle X2 in the ground truth map, rays can be emitted from the geometric center O of the real boundary W2′. The length of the line segment between each ray and the two intersection points of the perception boundary W2 and the real boundary W2′ is the offset of the perception boundary W2 along the ray direction. The predetermined offset may include the offset of the perception boundary W2 along at least one ray direction, or it may be calculated from the offset of the perception boundary W2 along multiple ray directions; this is not limited here.

[0120] like Figure 4 As shown, when determining the set offset, for example, four rays d1, d2, d3, and d4 (shown as dashed lines) can be emitted along the width and length directions of vehicle X2, with the geometric center O of the real boundary W2′ as the emission point. The offsets corresponding to the four rays are L1, L2, L3, and L4, respectively. Among them, the offset generated by the sensing boundary W2 along the direction of ray d1 is within the visible range of vehicle X1. That is, the intersection of ray d1 and the sensing boundary W2 is within the visible part of the sensing boundary W2. The offsets generated along the directions of rays d2 and d3 are blocked by vehicle X2, and the offset along the direction of ray d4 is blocked by column W1. Therefore, in Figure 4 In the example, the index value of the first local index can be taken as the offset L1 in the direction of ray d1.

[0121] Of course, what those skilled in the art can understand is that, Figure 4 In the example, more rays can be emitted from the geometric center O. If the intersections of multiple rays with the perception boundary W2 are all within the visible part of the perception boundary, the average or maximum value of the offsets corresponding to these rays can be calculated as the set offset, which is not limited here.

[0122] Furthermore, the traversable area is used by the automatic control system to determine the target entry area for a mobile task. The outer region of the target entry area is the necessary passageway for the automatic control system to move the mobile device from other areas to the target entry area. Therefore, the automatic control system's ability to perceive the traversable portion of the outer region surrounding the target entry area is equally crucial for safe driving. In some examples, the first type of indicator may include a second local indicator used to detect the accuracy of the boundary of the target entry area within the traversable area. Accordingly, the detection result includes the indicator value of the automatic control system for the second local indicator.

[0123] In this example, step S330, which compares the perception map and the truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene, may include the following steps: based on the perception map, determining the target entry area located within a second set range of the mobile device; and, for each target entry area, comparing the passable portion of the truth map and the perception map in the extension area of ​​the target entry area to obtain the indicator value of the automatic control system for the second local indicator.

[0124] In this example, the second set range can be represented by a distance relative to the mobile device, or by a square area of ​​a certain size with the mobile device as the geometric center. The distance or the size of the square area can be set according to the target scene or mobile task, and is not limited here.

[0125] In this example, the target entry area is related to the target scene or movement task. For example, when performing a parking task in a parking lot, the target entry area is an empty parking space. Or, when performing a lane-changing task on a road, the target entry area is an empty space in the adjacent lane.

[0126] For each target entry area, the outer region of the target entry area is determined in the perception map. This outer region can be obtained by offsetting the boundary of at least one side of the target entry area outward by a set distance.

[0127] Figure 5 The diagram illustrates a portion of a perception map for a parking task, generated by the automatic control system based on scene data collected by vehicle X1. Figure 5 The perception map indicates that there is an empty parking space C1 to the left front of vehicle X1, and this empty parking space C1 is the target entry area. When determining the index value of the automatic control system for the second local index, the left and right boundaries of the empty parking space C1 can be offset outward by a set distance, for example, 30cm-50cm, to obtain the outer extension area F1 of the empty parking space C1.

[0128] In this example, the Intersection over Union (IoU) ratio of the walkable portions of each extended region in the perceived map and the ground truth map can be calculated, and the average or maximum IoU ratio of the extended regions of all target-entering areas can be used as the index value of the second local index. Those skilled in the art should understand that the index value of the second local index can also be obtained by calculating the precision, recall, or accuracy of the perceived walkable portions in the extended region; this is not limited to this method.

[0129] Still with Figure 5 For example, for an extension region F1, let the passable part of the extension region F1 in the perception map be the first region A, and the passable part of the extension region F1 in the truth map be the second region B. Then the intersection-union ratio of the extension region F1 can be expressed as: the area of ​​the intersection of the first region A and the second region B divided by the area of ​​the union of the first region A and the second region B.

[0130] In some examples, regarding the first type of indicator, a first global indicator can also be set to detect the overall contour difference of the passable area, so as to specifically detect the perception accuracy of the automatic control system for the part affecting the movement decision. In this example, the detection result includes the indicator value of the automatic control system for the first global indicator. In this example, step S330, which compares the perception map and the ground truth map based on the set of indicators to obtain the detection result reflecting the scene perception capability of the automatic control system for the target scene, may include the following steps: emitting a set of sampling rays from the position of the mobile device in the perception map to the periphery, collecting the contour deviation value of the first contour of the passable area indicated by the perception map relative to the second contour of the passable area indicated by the ground truth map through the sampling rays; and determining the indicator value of the automatic control system for the first global indicator based on the contour deviation value collected by the set of sampling rays.

[0131] In this example, the contour deviation value between the first contour and the second contour can be determined based on the first intersection point of the sampling ray with the first contour and the second intersection point with the second contour. Specifically, the contour deviation value can be determined based on the line segment length of the same sampling ray between the first and second intersection points, or it can be determined based on the position coordinates of multiple first intersection points and multiple second intersection points acquired by a set of sampling rays.

[0132] When determining the contour deviation value between the first contour and the second contour based on the position coordinates of multiple first intersection points and multiple second intersection points, the minimum matching distance between the multiple first intersection points and the multiple second intersection points can be calculated. For example, the first intersection point can be matched with multiple second intersection points respectively to obtain the distance between the first intersection point and each second intersection point, and the minimum matching distance among them can be used as the contour deviation value based on the first intersection point, thereby obtaining multiple contour deviation values ​​based on the multiple first intersection points.

[0133] See this example. Figure 6 It can emit a set of sampling rays in four directions—forward, backward, left, and right—from vehicle X1 as the emission point. Figure 6 A portion of the sampling rays is schematically shown; the density of the sampling rays can be set as needed. Here, a single emission point can be selected on the vehicle to emit the sampling rays, for example, the geometric center of the vehicle can be used as the emission point. Alternatively, multiple emission points can be selected on the vehicle to emit the sampling rays, such as setting multiple emission points along the outline of the vehicle. No limitation is imposed here.

[0134] In this example, to improve the accuracy of sampling, valid samples can be selected by setting a sampling threshold. For example, samples with contour deviation values ​​within the sampling threshold range are set as valid samples, and the index values ​​of the first set of all indicators are determined based on the valid samples, while samples with contour deviation values ​​exceeding the sampling threshold are considered abnormal samples and are discarded.

[0135] Additionally, in some examples, at least one global metric can be set as a second type of metric to examine the overall detection performance of the automatic control system for passable areas. These global metrics can include at least one of the following: Intersection over Union (IoU), Precision, Recall, and Accuracy for passable area detection. This can be determined by comparing passable areas in the perception map with passable areas in the ground truth map.

[0136] To facilitate the calculation of the above indicators, the perception map and the ground truth map can be gridded. That is, the map is divided into multiple grids of a set size, with each grid representing a pixel in the map. Using the grid as the smallest unit for indicator calculation helps improve detection efficiency.

[0137] See Figure 7Taking the global metrics, which are classified as the second type of metric, as an example, we can statistically analyze the following: the first pixel TN, which is impassable in both the ground truth map (GT) and the perception map (PE); the second pixel TP, which is impassable in both the ground truth map and the perception map; the third pixel FP, which is impassable in the ground truth map but permissible in the perception map; and the fourth pixel FN, which is permissible in the ground truth map but impassable in the perception map. Based on these statistics, we can easily calculate the metric values ​​related to intersection-union ratio (IU), precision, recall, and accuracy.

[0138] Regarding the map comparison in step S330, in the examples where the perceptual map includes multiple perceptual map segments and the ground truth map includes multiple ground truth map segments, the perceptual map segments and ground truth map segments corresponding to the same frame data can be compared separately based on a set of indicators to obtain the detection result. That is, among the multiple perceptual map segments and multiple ground truth map segments, multiple map segment pairs are formed based on the corresponding data frames. In step S330, each map segment pair is compared separately to obtain the detection result.

[0139] In such an example, the detection results may include single-frame detection results for each indicator for each map fragment pair, and / or, the average value of each indicator based on multiple single-frame detection results. This multidimensionality of the detection results is beneficial for problem localization, enabling targeted optimization of the automatic control system to meet the usage requirements of different or specific scenarios.

[0140] As described in steps S310 to S330, the method of this embodiment compares the perception map with the ground truth map based at least on indicators associated with the location of the mobile device in the perception map, thereby detecting the scene perception capability of the automatic control system. Since the location-associated indicators can provide location-based perception bias, the method of this embodiment can determine the safety of the motion control performed by the automatic control system in the target scene to complete the motion task, thereby improving the effectiveness of detection and further system optimization.

[0141] <Second Embodiment>

[0142] In the second embodiment, the automatic control system can be targeted to detect scene perception capabilities based on the different mobile tasks, thereby improving the automatic control system's perception capabilities for various scenarios through the detection results of different mobile tasks.

[0143] In this embodiment, the target scene corresponds to a first movement task in the set of tasks. For example, the target scene is a parking lot, and the first movement task is a parking task. The set of tasks also includes a second movement task, such as a lane-changing task or a driving task within a lane. In this embodiment, the set of indicators is a first set of indicators that matches the first movement task, and the first set of indicators that matches the second movement task is different from the second set of indicators that matches the first movement task.

[0144] In this embodiment, different indicator sets may include different indicator types. For example, the first indicator set includes global indicators and local indicators, while the second indicator set only contains global indicators. Alternatively, both the first and second indicator sets may include global indicators and local indicators, but they may have different types of local indicators.

[0145] In this embodiment, different indicator sets can also include: the indicator types are the same, but the objects targeted by the indicators are different. For example, both the first indicator set and the second indicator set include the first local indicators mentioned above, but the target obstacles targeted by the first local indicators in the first indicator set include vehicles and pillars, while the target obstacles targeted by the first local indicators in the second indicator set include vehicles and pedestrians, etc.

[0146] In some examples, for driving tasks that do not require special accuracy in perceiving nearby areas, detection can be performed solely using global indicators to improve the detection efficiency of the automatic control system. In other words, the task set can include only moving tasks that match global indicators, such as driving tasks within a lane. In this example, moving tasks in the task set can be divided into two categories based on whether the indicator set matched by the moving task includes local indicators. The first category of tasks matches indicators that include at least local indicators, while the second category matches indicators that include only global indicators.

[0147] Since global indicators are used to detect the overall perception effect of automatic control systems and can be adapted to all movement tasks, in some examples, different movement tasks in the task set can share global indicators. That is to say, the difference in indicator sets is mainly reflected in the setting of local indicators. By setting matching local indicators for movement tasks, the objects that the automatic control system should focus on when performing the corresponding movement tasks can be detected in a targeted manner.

[0148] In such an example, such as Figure 8 As shown, a method for detecting an automatic control system may include the following steps:

[0149] Step S810: Obtain the perception map of the target scene generated by the automatic control system based on the target scene data; wherein, the target scene data is collected from the target scene by the mobile device.

[0150] Step S820: Obtain the ground truth map of the target scene based on the ground truth annotation of the target scene data.

[0151] Step S8301: Based on the global index comparison between the perception map and the truth map, the first detection result for the automatic control system is obtained.

[0152] Step S8302: Determine whether the first movement task is a first type of task. If not, it means that the first movement task is a second type of task. Execute step S8303. If yes, execute step S8304.

[0153] Step S8303: Output the first detection result as the final detection result.

[0154] Step S8304: Determine the local index that matches the first moving task, and compare the perception map and the ground truth map based on the determined local index to obtain the second detection result for the automatic control system.

[0155] Step S8305: Output the first detection result and the second detection result as the final detection result.

[0156] In this example, steps S8301 and S8302 are decoupled, so that steps S8301 and S8302 can be executed synchronously, which is beneficial to improving detection efficiency.

[0157] <Third Embodiment>

[0158] This embodiment provides an optimization method for an automatic control system, which may include the following steps:

[0159] Step S310: Obtain the perception map of the target scene generated by the automatic control system based on the target scene data; wherein, the target scene data is collected by the mobile device in the target scene.

[0160] Step S320: Obtain the truth map of the target scene based on the truth labeling of the target scene data.

[0161] Step S330: Based on the set of indicators, compare the perception map and the truth map to obtain the detection results reflecting the scene perception capability of the automatic control system for the target scene; wherein, the indicator set includes indicators associated with the position of the mobile device in the perception map.

[0162] Step S340: Optimize the automatic control system based on the detection result to obtain the optimized automatic control system.

[0163] The optimization in this embodiment can be an adjustment to any part of the automatic control system. This optimization includes adjusting the parameter values ​​of the automatic control system. The optimization may also include adjusting the system structure of the automatic control system, such as adjusting the model composition of the automatic control system, or adjusting the network structure of at least one model in the automatic control system, etc., without limitation.

[0164] In this embodiment, since the detection results can reflect the deficiencies of the automatic control system in scene perception, the automatic control system can be optimized to overcome these deficiencies based on the detection results, thereby improving the performance of the automatic control system in a targeted manner, especially by combining the index values ​​of position-related indicators to improve the safety of its mobile task execution.

[0165] <Fourth Embodiment>

[0166] This embodiment provides a control method for a mobile device, which is implemented by a first automatic control system configured for the first mobile device, such as... Figure 9 As shown, the method may include the following steps S910 and S920:

[0167] Step S910: Obtain the first scene data collected by the first mobile device.

[0168] The first mobile device can collect first scene data through at least one sensor it carries, which will not be elaborated here.

[0169] Step S920: Based on the first scene data, control the first mobile device to move.

[0170] In this embodiment, the first automatic control system can determine the passable area in the corresponding scene based on the first scene data, and then plan a movement path based on the determined passable area and the movement task, and control the first mobile device to move according to the movement path to complete the movement task.

[0171] In this embodiment, the first automatic control system is obtained by adjusting the parameter values ​​or system structure of the second automatic control system; that is, the first automatic control system is obtained by optimizing the second automatic control system. The parameter values ​​here can be the model parameter values ​​of the automatic control system. The system structure can include the network structure and / or model composition of the model, etc., and is not limited here.

[0172] The above adjustments are based on the detection results of the scene perception capability of the second automatic control system for at least one target scene. The detection results are obtained by comparing the perception map and the truth map for an index set. The index set includes indicators associated with the position of the mobile device in the perception map, which is generated by the second automatic control system based on the second scene data collected by the second mobile device in the target scene. The truth map is obtained by labeling the second scene data with truth values.

[0173] In this embodiment, the first automatic control system can be a general automatic control system obtained through multi-scenario detection and optimization.

[0174] In this embodiment, the first automatic control system can also be an automatic control system dedicated to certain scenarios. That is, at least two automatic control systems can be configured for the mobile device. The mobile device can call the automatic control system adapted to the current scenario to perform the movement task, so that the mobile device has excellent mobility in different scenarios. In order to achieve the lightweighting of the automatic control system, these automatic control systems can share at least some models and some parameter values. In the optimization, only specific parts of the system need to be adjusted for different scenarios to improve the optimization efficiency.

[0175] In this embodiment, the first mobile device and the second mobile device can be the same mobile device or different mobile devices, and no limitation is made here.

[0176] The system optimization in this embodiment can be performed offline or online, and no limitation is made here.

[0177] <Fifth Embodiment>

[0178] This embodiment provides a chip capable of implementing the methods of the embodiments of this disclosure, such as... Figure 10 As shown, the chip 1000 includes a storage unit 1020 and a processing unit 1010. The storage unit 1020 is used to store a computer program. The processing unit 1010 is configured to implement the method according to any embodiment of this disclosure when executing the computer program stored in the storage unit.

[0179] Chip 1000 can be a processor chip with data processing capabilities.

[0180] Chip 1000 can be a processor chip used in Automated Driving Control Units (ADCUs).

[0181] Chip 1000 can be a system-on-a-chip (SoC) that integrates multiple processors, multiple memories, I / O interfaces, etc., to achieve miniaturized controller design.

[0182] <Sixth Embodiment>

[0183] This embodiment provides an electronic device, such as... Figure 11 As shown, the electronic device 1100 includes a memory 1102 and a processor 1101. The memory 1102 is used to store a computer program executed by the processor 1101. The processor 1101 is configured to implement the method according to any embodiment of the present disclosure when executing the computer program stored in the memory 1102.

[0184] The electronic device can be a controller for a mobile device, such as a vehicle's intelligent driving domain controller or a vehicle's central controller.

[0185] The electronic device can also be a server, etc., and there is no limitation here.

[0186] <Seventh Embodiment>

[0187] This embodiment provides a mobile device, which may include a chip 1000 according to the fifth embodiment or an electronic device 1100 according to the sixth embodiment. In this case, the electronic device 1100 may be a controller of the mobile device.

[0188] In some embodiments, such as Figure 12 As shown, the mobile device 1200 may also include a memory 1202 and a processor 1201. The memory 1202 is used to store a computer program executed by the processor 1201. The processor 1201 is configured to implement the method according to any embodiment of the present disclosure when executing the computer program stored in the memory 1202.

[0189] The mobile device 1200 can be, for example, a vehicle with intelligent driving capabilities, or an intelligent robot of any form, etc., without limitation.

[0190] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the methods in the foregoing embodiments of this disclosure. Optionally, the computer-readable storage medium may be a non-transitory storage medium, but is not limited thereto; it may also be a temporary storage medium.

[0191] This disclosure also provides a computer program product, which may include a computer program that, when executed by a processor, can implement any of the methods described in the foregoing embodiments of this disclosure.

[0192] The above embodiments focus on illustrating the differences from other embodiments. The same or similar parts between different embodiments can be referred to each other.

[0193] This disclosure may be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement any of the methods in the foregoing embodiments of this disclosure.

[0194] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example, but not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0195] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0196] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0197] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0198] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0199] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions that execute on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0200] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. It should be noted that implementation in hardware, implementation in software, and implementation using a combination of software and hardware are all equivalent.

[0201] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, and are not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein. The scope of this disclosure is defined by the appended claims.

Claims

1. A method for detecting an automatic control system, characterized in that, include: An automatic control system acquires a perception map of a target scene generated based on target scene data; wherein the target scene data is collected by a mobile device within the target scene. Obtain a truth map of the target scene based on truth labeling of the target scene data; By comparing the perception map and the ground truth map based on a set of defined indicators, a detection result reflecting the scene perception capability of the automatic control system for the target scene is obtained; wherein, the set of defined indicators includes indicators associated with the position of the mobile device in the perception map, and the indicators reflect the perception capability of the automatic control system for passable areas in the target scene.

2. The method according to claim 1, characterized in that, The set of defined indicators includes at least one of local indicators associated with the location and global indicators associated with the location; wherein, Local metrics associated with the location include: metrics for detecting the scene perception capability based on a local region of the perception map, wherein the local region is a region determined based on the location of the mobile device in the perception map; The global metrics associated with the location include: metrics for the scene perception capability based on the entire area of ​​the perception map and objects associated with the location.

3. The method according to claim 1, characterized in that, The perception map indicates the passable area of ​​the mobile device in the target scene as perceived by the automatic control system; the truth map indicates the passable area of ​​the mobile device in the target scene determined based on the truth annotation.

4. The method according to claim 1, characterized in that, The set of defined indicators is a first set of indicators that matches the first mobile task corresponding to the target scenario; The first mobile task is a mobile task in a set of defined tasks. The set of defined tasks also includes a second mobile task. The second set of indicators that matches the second mobile task is different from the first set of indicators that matches the first mobile task.

5. The method according to claim 1, characterized in that, The target scene data includes multiple frames of data, with the same frame corresponding to the same timestamp and different frames corresponding to different timestamps; the perception map includes multiple perception map segments that correspond one-to-one with the multiple frames of data, and the truth map includes multiple truth map segments that correspond one-to-one with the multiple frames of data. The step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result representing the scene perception capability of the automatic control system for the target scene includes: Based on a set of indicators, the perceptual map fragments and ground truth map fragments of corresponding data in the same frame are compared to obtain the detection results.

6. The method according to claim 4, characterized in that, When the first mobile task is a first type of task, the set of indicators includes local indicators of the scene perception capability based on the local area detection of the perception map. When the first mobile task is a second type of task, the indicators in the indicator set are global indicators of the scene perception capability based on the entire area of ​​the perception map.

7. The method according to claim 6, characterized in that, The global indicators are indicators shared by different mobile tasks. The step of comparing the perception map and the ground truth map based on a set of indicators to obtain detection results reflecting the scene perception capability of the automatic control system for the target scene includes: Based on the global index, the perception map and the truth map are compared to obtain the first detection result for the automatic control system. If the first mobile task is the second type of task, the first detection result is output as the detection result. If the first mobile task is a first type of task, determine the local metrics that match the first mobile task; Based on the determined local indicators, the perception map and the truth map are compared to obtain a second detection result for the automatic control system. The first detection result and the second detection result are output as the detection result.

8. The method according to claim 2, characterized in that, The global indicators associated with the location include a first global indicator for detecting contour differences in traversable areas, the detection result including the indicator value of the automatic control system for the first global indicator; the step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes: A set of sampling rays is emitted from the position of the mobile device in the sensing map to the periphery, and the following are collected through the sampling rays: the contour deviation value of a first contour of a passable area indicated by the sensing map relative to a second contour of a passable area indicated by the ground truth map; wherein the contour deviation value is determined based on a first intersection point of the sampling ray with the first contour and a second intersection point of the sampling ray with the second contour; Based on the contour deviation values ​​collected by the set of sampling rays, the index value of the automatic control system for the first global index is determined.

9. The method according to claim 2, characterized in that, The local indicators associated with the location include a first local indicator for detecting obstacle boundary errors, and the detection result includes the indicator value of the automatic control system for the first local indicator; the step of comparing the perception map and the ground truth map based on a set of indicators to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes: Based on the perception map, target obstacles within a first predetermined range of the mobile device are identified; Based on each target obstacle, using the true boundary of the target obstacle in the truth map as a reference, the set offset of the perceived boundary of the target obstacle in the perception map relative to the true boundary is determined, and used as the index value of the automatic control system for the first local index. Wherein, the set offset is the offset of the visible portion of the perceived boundary within the visible range of the mobile device relative to the real boundary.

10. The method according to claim 2, characterized in that, The local metrics associated with the location include a second local metric for detecting the accuracy of the boundary of the target entering the traversable area, and the detection result includes the metric value of the automatic control system for the second local metric; the step of comparing the perception map and the ground truth map based on a set of metrics to obtain a detection result reflecting the scene perception capability of the automatic control system for the target scene includes: Based on the perception map, the target entry area located within a second predetermined range of the mobile device is determined; For each target entry area, the passable portion of the ground truth map and the perception map in the extended region of the target entry area is compared to obtain the index value of the automatic control system for the second local index.

11. A control method for a mobile device, said control method being implemented by a first automatic control system, characterized in that, The method includes: Acquire the first scene data collected by the first mobile device; Based on the first scene data, control the movement of the first mobile device; The first automatic control system obtains the values ​​of the parameters of the second automatic control system by adjusting the parameters of the second automatic control system. The adjustment is based on the detection results of the scene perception capability of the second automatic control system for at least one target scene. The detection results are obtained by comparing the perception map and the truth map based on the index set. The index set includes indicators associated with the position of the mobile device in the perception map. The indicators reflect the perception capability of the automatic control system for traversable areas in the target scene. The perception map is generated by the second automatic control system based on the second scene data collected by the second mobile device in the target scene. The truth map is obtained by labeling the second scene data with truth values.

12. A chip, characterized in that, The chip includes a storage unit and a processing unit, wherein the storage unit stores a computer program and the processing unit executes the computer program stored in the storage unit to implement the method as described in any one of claims 1 to 11.

13. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store computer instructions, and the processor being used to retrieve the computer instructions from the memory to perform the method as described in any one of claims 1 to 11.

14. A mobile device, characterized in that, Includes the chip of claim 12, or the electronic device of claim 13; or, The mobile device includes a memory and a processor, the memory storing computer instructions and the processor retrieving the computer instructions from the memory to perform the method as described in any one of claims 1 to 11.

15. A computer-readable storage medium, characterized in that, It stores a computer program thereon, which, when executed by a processor, implements the method as described in any one of claims 1 to 11.

Images